The goal of this project is to use MRI and CSF biomarkers to identify in cognitively normal (NL) subjects, the earliest clinically detectable brain changes of Alzheimer's disease (AD). Neuropathology studies show, in mild cognitive impairment (MCI) and some NL individuals, neurofibrillary and plaque pathology that is associated with neuronal and volume losses in the entorhinal cortex (EC) and hippocampus. Our recent in vivo MRI and PET studies of the EC show that volume and metabolism reductions predict the conversion of NL to future MCI. However, in vivo imaged changes lack disease specificity. Our most recent studies and those of our collaborators show that tangle-related abnormal tau proteins, hyper-phosphorylated at threonine 231 (P-tau231), are: 1) elevated in the CSF in MCI; 2) useful in predicting further cognitive decline; and 3) diagnostically specific for AD. Our cross-sectional diagnosis and longitudinal data demonstrate that P-tau231 significantly adds to the MRI hippocampal volume in the diagnosis of MCI. Moreover, we recently published an MRI-based technique that improves detection of longitudinal P-tau231 increases in MCI by controlling for progressive ventricular CSF volume dilution. Our pilot studies also show that amyloid beta 1-40 (Abeta40) levels are elevated in MCI. Overall, these findings suggest that a sensitive and specific recognition of AD in the normal stages of cognition is close at hand. We plan to complete over 5-years a longitudinal MRI and CSF study consisting of a baseline and two 18-month follow-up exams on 80 elderly NL subjects currently active in a longitudinal MRI study of aging and memory. To enrich the sample with individuals at increased risk for progressive cognitive decline, 50 of the subjects will be selected on the basis of subjective memory complaints. Longitudinally studied FTD patients are available to examine specificity. We will use quality controlled and standardized protocols at each observation to collect norm-referenced neuropsychological data; high-resolution MRI; and CSF. Our major hypothesis is that CSF P-tau231 measurement improves the accuracy (specificity) of MRI EC atrophy and CSF Abeta measurements in prediction of the conversion from NL to MCI. All the required imaging and clinical components for this study are active. Ample numbers of subjects are available, and we already completed the proposed three evaluations on 8 eligible NL subjects. There is adequate statistical power for testing the hypotheses.